Cryptic biodiversity of freshwater fish species in Bangladesh

PLOS One | https://doi.org/10.1371/journal.pone.0318982 April 28, 2025 8 / 18 PLOS One Cryptic biodiversity of freshwater fish Phylogenetical analysis The resulting COI gene nucleotide sequences were matched with additional COI sequences of Bangladeshi fish from INSD [14] and the MAFFT (Multiple Alignment using Fast Fourier Transform) v. 7.427 with the L-INS-I option [21]. Phylogenetic trees were inferred using maximum likelihood (ML) and Bayesian inference (BI) techniques. IQ-TREE v. 1.6.12 was used to estimate nucleotide substitution models for ML and BI analyses using AICc [22]. The IQ-TREE was used to determine the ML phylogeny, and 1000 pseudo replicates were used for an ultra-fast bootstrapping (BS). MrBayes v. 3.2.6 was used to obtain the Bayesian poste- rior probabilities and BI tree [23]. A tree sample was obtained after every 1000 generations, throughout two separate runs of four Markov chains totaling 100 million generations. Tracer v. 1.7.1 [24] was used to examine the parameter estimates and convergence, and the first 10% trees were eliminated in light of the findings. MEGA11’spairwise-deletion option was used for computing pairwise comparisons of the uncorrected p-distance [25] (S1 Table). Species delimitation analyses and morphology measurement Species delimitation investigations were conducted using Assemble Species by Automatic Par- titioning (ASAP) [26] based on 712 COI sequences used in phylogenetic analysis (S2 Table). The ASAP analysis was conducted using the online ASAP version (https://bioinfo.mnhn.

fr/abi/public/asap/asapweb.html) using the uncorrected pairwise distances with the default parameters. We looked at morphological parameters under both a microscope and a magnify- ing glass. We only took morphometric measurements of the chosen cryptic fish using a digital vernier caliper (Mitutoyo, accuracy ±0.02 mm). Ethics statement The specimens were obtained from either from markets or fishermen, or they were collected from the wild using a beach seine or hand net. The fish were then euthanized by immersion in buffered tricaine-methanesulphonate (MS 222) for thirty minutes after the animals stopped moving, in accordance with the guidelines outlined in the permits issued by Jamalpur Science and Technology University’s Ethics Committee (Letter no. 37.01.0044.064.05.001.24.553). Results Fish specimens were gathered nationwide from more than 28 locations (Table 1), and ana- lyzed the mitochondrial COI sequences of 124 individuals representing 121 species, 94 genera, 39 families, and 19 orders. Sequences with ≥1 nucleotide change were considered distinct haplotypes. Our COI alignment matrix consisted of 655 bp representing 119 haplotypes. Table 1 provides a summary of all haplotypes. Based on COI sequences, nearly identical topologies were observed in the ML (ln L = −59788.54) and BI (ln L = −60808.92) trees (S1 Fig). Uncorrected p-distance and species delimitation analysis demonstrated 243 groups with the best score. Most of the OTUs (operational taxonomy units) sequenced in this study were

assigned to groups with known species; however, some specimens formed their own unique groups. In particular, seven OTUs had to be assigned to different groups, despite the presence of the same species identified morphologically, indicating that these specimens were cryptic species (Table 2). Specifically, the specimen of Ailia sp. (Ailiidae) had been morphologically identified as Ailia coila, but genetically it was found to belong to a different species (Ailia sp.) compared to its nearest congener (COI divergence 8.2%) (Fig 1). This specimen was collected from the area adjacent to the Jamuna River (Sariakandi, Bogura). Our examined specimens of genus Awaous